sglang

ascend

[Bugfix] Fix Qwen3/DSV3/DSV3.2 model support (#11510 )

2025-10-16 15:14:09 +08:00

batch_invariant

Fix batch invariant ops (#11368 )

2025-10-10 20:49:08 -07:00

configs

Add test config yamls for Deepseek v3 (#5433 )

2025-04-20 17:28:52 -07:00

cpu

[CPU] Optimize FP16 decode_attention_cpu (#10652 )

2025-10-22 21:39:51 -07:00

entrypoints/http_server

[RL] Add test for /abort_request (#7626 )

2025-08-10 09:14:19 -07:00

ep

[2/N] Added the core structure of elastic EP and the eplb algorithm with faulty rank (#10606 )

2025-10-22 01:13:31 -07:00

function_call

Use jsonschema to constrain required or specific tool choice (#10550 )

2025-09-27 13:18:50 -04:00

hicache

Fix flaky hicache test with mooncake backend (#11953 )

2025-10-22 21:00:47 +08:00

layers/attention

Feature/nano v2 offline modelopt fp8 and nvfp4 (#12018 )

2025-10-23 11:16:46 -07:00

lora

adjust dynamic vs static outputs comparison in test_lora_update.py (#11884 )

2025-10-24 10:35:34 -07:00

models

Feature/nano v2 offline modelopt fp8 and nvfp4 (#12018 )

2025-10-23 11:16:46 -07:00

openai_server

Support Thinking Budget (via custom_logit_processor for OpenAI API) [Fix #6572 ] (#11416 )

2025-10-21 16:27:56 +08:00

quant

[2/N]Support DeepSeek-R1 w4a8 low latency deepep (#8464 )

2025-10-24 17:41:16 -07:00

rl

[minor] sync code on python/sglang/test/test_deterministic.py and improve ci tests (#11777 )

2025-10-17 14:25:22 -07:00

rotary_embedding

Support mrope triton kernel and add unit test (#11722 )

2025-10-20 11:51:07 +08:00

xpu

Init attention backend for Intel XPU (#10656 )

2025-10-21 11:41:28 +08:00

double-sparsity-config-Llama-3.1-8B-Instruct.json

Rename double sparsity config file (#2188 )

2024-11-25 17:12:08 -08:00

experiment_runner.py

Replace time.time() to time.perf_counter() for benchmarking. (#6178 )

2025-05-11 14:32:49 -07:00

kv_cache_scales_llama3_1_8b.json

Support FP8 E4M3 KV Cache (#2786 )

2025-01-12 21:17:11 -08:00

kv_cache_scales_llama3_8b.json

support e4m3 kvcache in qwen2 & add kv scaling facotr json (#2894 )

2025-01-18 11:43:22 +08:00

kv_cache_scales_qwen2_1_5b.json

support e4m3 kvcache in qwen2 & add kv scaling facotr json (#2894 )

2025-01-18 11:43:22 +08:00

parse_results.py

misc: parse bench_serving result as markdown table (#9377 )

2025-08-20 16:44:20 -07:00

run_suite.py

[2/N]Support DeepSeek-R1 w4a8 low latency deepep (#8464 )

2025-10-24 17:41:16 -07:00

test_abort.py

Support updating weights at once by stopping all requests (#6698 )

2025-07-02 22:26:06 -07:00

test_async_dynamic_batch_tokenizer.py

[Performance] Dynamic Batch Tokenizer (#9382 )

2025-09-14 01:56:04 +08:00

test_bench_one_batch.py

Add V2-lite model test (#7390 )

2025-07-03 22:25:50 -07:00

test_bench_serving.py

[ci] fix pp test (#11294 )

2025-10-07 14:20:04 +08:00

test_bnb.py

vlm: enforce pybase64 for image and str encode/decode (#10700 )

2025-10-21 19:05:32 +08:00

test_build_eagle_tree.py

move eagle draft post process to cuda graph (#11434 )

2025-10-14 22:50:53 +08:00

test_chunked_prefill.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_config_integration.py

chore: Initial support for input config files (#10534 )

2025-09-24 14:45:52 -07:00

test_cpp_radix_cache.py

[Feature] Radix Tree in C++ (#7369 )

2025-08-02 19:50:14 -07:00

test_create_kvindices.py

fix: revert #8593 (#9581 )

2025-08-25 01:29:06 -07:00

test_custom_allreduce.py

[AMD] Add unit-test-sgl-kernel-amd to AMD CI (#7539 )

2025-06-29 15:50:09 -07:00

test_cutedsl_flashinfer_8gpu.py

[NVIDIA] [3/N] Nvfp4 Masked Gemm: Add flashinfer grouped_gemm_nt_masked (#9199 )

2025-09-11 20:18:43 -07:00

test_data_parallelism.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_deepseek_v3_basic.py

reverse the amd ci test back to 1200s and split the 8-gpu deepseek job into two. (#11238 )

2025-10-06 19:27:57 -04:00

test_deepseek_v3_fp4_4gpu.py

fix: update dsv3 fp4 ut (#10584 )

2025-09-17 14:34:58 -07:00

test_deepseek_v3_mtp.py

reverse the amd ci test back to 1200s and split the 8-gpu deepseek job into two. (#11238 )

2025-10-06 19:27:57 -04:00

test_deepseek_v32_basic.py

[CI] Add CI test for DeepSeek V3.2 MTP (#11835 )

2025-10-19 17:00:25 -07:00

test_deepseek_v32_mtp.py

[CI] Add CI test for DeepSeek V3.2 MTP (#11835 )

2025-10-19 17:00:25 -07:00

test_deterministic.py

convert test_deterministic into unit tests (#11095 )

2025-10-06 20:33:11 -07:00

test_disaggregation_basic.py

Reorder PD disagg CI tests (#11438 )

2025-10-10 17:56:49 -07:00

test_disaggregation_different_tp.py

[ci]use H20 to run disaggregation test (#11543 )

2025-10-16 11:42:42 -07:00

test_disaggregation_dp_attention.py

Move deep gemm related arguments to sglang.srt.environ (#11547 )

2025-10-14 00:34:35 +08:00

test_disaggregation_hybrid_attention.py

[PD] Add PD support for hybrid model (Qwen3-Next, DeepSeek V3.2 Exp) (#10912 )

2025-10-15 18:59:14 -07:00

test_disaggregation_pp.py

[ci]use H20 to run disaggregation test (#11543 )

2025-10-16 11:42:42 -07:00

test_double_sparsity.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_dp_attention.py

Remove dp balance metadata and minimul token balance. (#11170 )

2025-10-03 01:48:15 +08:00

test_eagle_infer_a.py

Tiny move files to utils folder (#11166 )

2025-10-03 22:40:06 +08:00

test_eagle_infer_b.py

[EAGLE] Refactor code for page size > 1 & more simplifications (#7213 )

2025-06-16 03:04:29 -07:00

test_eagle_infer_beta.py

[Test] Add basic matched stop for beta eagle (#11833 )

2025-10-20 01:17:00 +08:00

test_ebnf_constrained.py

ci: Fix test_ebnf_generate_all_optional_function_params (#7288 )

2025-06-17 16:39:42 -07:00

test_eval_accuracy_large.py

Fix GPU OOM (#6564 )

2025-05-24 16:38:39 -07:00

test_eval_fp8_accuracy.py

[CI] Fix ci tests (#5769 )

2025-04-27 07:18:10 -07:00

test_expert_distribution.py

Move deep gemm related arguments to sglang.srt.environ (#11547 )

2025-10-14 00:34:35 +08:00

test_expert_location_updater.py

Move files related to EPLB (#7580 )

2025-06-29 15:39:38 -07:00

test_fa3.py

Move deep gemm related arguments to sglang.srt.environ (#11547 )

2025-10-14 00:34:35 +08:00

test_fim_completion.py

Tiny move files to utils folder (#11166 )

2025-10-03 22:40:06 +08:00

test_flash_attention_4.py

[NVIDIA] FA3/FA4 Fix (#11606 )

2025-10-19 17:10:10 -07:00

test_flashmla.py

[fix]enable flashmla when using draft model P/D attention select (#11012 )

2025-10-04 20:59:34 +08:00

test_forward_split_prefill.py

Separate allocation logic from scheduler (#11313 )

2025-10-10 17:38:54 -07:00

test_function_call_parser.py

Tiny move files to utils folder (#11166 )

2025-10-03 22:40:06 +08:00

test_fused_moe.py

[6/N] MoE Refactor: Cleanup MoE-related configs (#8849 )

2025-08-14 21:14:53 -07:00

test_get_weights_by_name.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_gguf.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_gpt_oss_1gpu.py

Fix CI (#9013 )

2025-08-09 16:00:10 -07:00

test_gpt_oss_4gpu.py

feat: add gpt oss b200 ci (#9988 )

2025-09-03 17:26:38 -07:00

test_gpt_oss_common.py

Add tests to AMD CI for MI35x (#9662 )

2025-09-10 12:50:05 -07:00

test_gptqmodel_dynamic.py

Depreate global_server_args_dict (#11528 )

2025-10-13 19:34:43 +08:00

test_harmony_parser.py

Move parsers under a single folder (#9912 )

2025-09-02 18:25:04 -07:00

test_health_check.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_hidden_states.py

Open AI API hidden states (#6716 )

2025-06-10 14:37:29 -07:00

test_hybrid_attn_backend.py

Move deep gemm related arguments to sglang.srt.environ (#11547 )

2025-10-14 00:34:35 +08:00

test_input_embeddings.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_io_struct.py

fix: Handles input_embeds in GenerateReqInput when n>1 (#7830 )

2025-07-08 14:00:42 -07:00

test_jinja_template_utils.py

Move parsers under a single folder (#9912 )

2025-09-02 18:25:04 -07:00

test_kv_events.py

feat: add dp-rank to KV events (#6852 )

2025-06-04 15:29:34 -07:00

test_load_weights_from_remote_instance.py

Support loading weights from remote instance (#8215 )

2025-09-12 17:40:22 +08:00

test_local_attn.py

Simplify FA3 tests (#5779 )

2025-04-27 01:30:17 -07:00

test_logprobs.py

[logprobs] Enable local deterministic logrprobs testing with strict threshold (#10994 )

2025-10-19 13:30:39 -07:00

test_mamba_unittest.py

[PD] Add PD support for hybrid model (Qwen3-Next, DeepSeek V3.2 Exp) (#10912 )

2025-10-15 18:59:14 -07:00

test_metrics_utils.py

Rename customer label -> custom label (#10899 )

2025-09-25 16:19:53 -07:00

test_metrics.py

Improve structured outputs: fix race condition, server crash, metrics and style (#6188 )

2025-05-11 08:36:16 -07:00

test_mla_deepseek_v3.py

Use more general heuristics to set the default value of --mem-fraction-static (#10975 )

2025-09-29 10:11:03 -07:00

test_mla_flashinfer.py

fix: revert #8593 (#9581 )

2025-08-25 01:29:06 -07:00

test_mla_fp8.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_mla_int8_deepseek_v3.py

Fix unit tests (#11503 )

2025-10-12 07:45:57 -07:00

test_mla_tp.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_mla.py

Improve profiler and integrate profiler in bench_one_batch_server (#6787 )

2025-05-31 15:53:55 -07:00

test_modelopt_export.py

Enable native ModelOpt quantization support (3/3) (#10154 )

2025-10-21 21:44:29 -07:00

test_modelopt_fp8kvcache.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_modelopt_loader.py

Enable native ModelOpt quantization support (3/3) (#10154 )

2025-10-21 21:44:29 -07:00

test_modelopt.py

[CI] Fix ci tests (#5769 )

2025-04-27 07:18:10 -07:00

test_models_from_modelscope.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_moe_eval_accuracy_large.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_mscclpp.py

support 1 shot allreduce in 1-node and 2-node using mscclpp (#6277 )

2025-06-04 22:11:24 -07:00

test_multi_instance_release_memory_occupation.py

Fix mem fraction static for nightly tests (#11076 )

2025-09-29 12:57:41 -07:00

test_multi_tokenizer.py

Support Multi Process Tokenizer Manager(#6555 ) (#8964 )

2025-09-01 01:00:13 -07:00

test_ngram_speculative_decoding.py

Move deep gemm related arguments to sglang.srt.environ (#11547 )

2025-10-14 00:34:35 +08:00

test_nightly_gsm8k_eval_amd.py

ci: refactor nightly test (#10495 )

2025-09-26 15:24:30 -07:00

test_nightly_text_models_gsm8k_eval.py

ci: unify the model launch method of nightly ci (#11230 )

2025-10-07 18:13:14 -07:00

test_nightly_text_models_perf.py

ci: improve nightly-ci (#11385 )

2025-10-12 21:19:34 -07:00

test_nightly_vlms_mmmu_eval.py

ci: fix night-ci with push retry mechanism (#11765 )

2025-10-23 11:31:05 -07:00

test_nightly_vlms_perf.py

ci: improve nightly-ci (#11385 )

2025-10-12 21:19:34 -07:00

test_no_chunked_prefill.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_no_overlap_scheduler.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_original_logprobs.py

Sync code and test CI; rename some env vars (#11686 )

2025-10-15 18:37:03 -07:00

test_page_size.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_patch_torch.py

Tiny move files to utils folder (#11166 )

2025-10-03 22:40:06 +08:00

test_penalty.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_piecewise_cuda_graph.py

support more model in piecewise cuda graph (#11745 )

2025-10-24 10:31:39 +08:00

test_pp_single_node.py

[Fix] Compatibility between DP attention and pipeline parallelism (#10100 )

2025-09-06 01:34:10 -07:00

test_priority_scheduling.py

feat: add priority based scheduling with priority based request acceptance and preemption (#8746 )

2025-09-16 17:10:10 -07:00

test_profile_merger_http_api.py

[FEATURE] Add Profile Trace Merger for Distributed Traces (#11413 )

2025-10-14 09:20:17 +08:00

test_profile_merger.py

[FEATURE] Add Profile Trace Merger for Distributed Traces (#11413 )

2025-10-14 09:20:17 +08:00

test_pytorch_sampling_backend.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_quick_allreduce.py

[Feature] Integrate quick allreduce and select the best allreduce implementation (#6619 )

2025-07-24 20:48:42 -07:00

test_radix_attention.py

Improve profiler and integrate profiler in bench_one_batch_server (#6787 )

2025-05-31 15:53:55 -07:00

test_radix_cache_unit.py

Fix eagle radix cache (#10846 )

2025-09-30 22:59:20 +08:00

test_reasoning_parser.py

Move parsers under a single folder (#9912 )

2025-09-02 18:25:04 -07:00

test_regex_constrained.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_release_memory_occupation.py

[Feature] Option to save model weights to CPU when memory saver mode is enabled (#10873 )

2025-10-03 16:48:19 +08:00

test_request_queue_validation.py

feat: add priority based scheduling with priority based request acceptance and preemption (#8746 )

2025-09-16 17:10:10 -07:00

test_retract_decode.py

[Fix] memory leak by overlap + retract (#11981 )

2025-10-23 22:59:23 +08:00

test_rope_rocm.py

Revert "[ROCm] Remove vLLM rope dependency & use AITER impl" (#12028 )

2025-10-23 12:42:59 -07:00

test_sagemaker_server.py

Tiny move files to utils folder (#11166 )

2025-10-03 22:40:06 +08:00

test_schedule_policy.py

feat: add priority based scheduling with priority based request acceptance and preemption (#8746 )

2025-09-16 17:10:10 -07:00

test_score_api.py

[Generative Score API] Multi-Item scoring with custom attention mask. (#10979 )

2025-10-08 18:47:32 -07:00

test_server_args.py

Improve dp attention port assignment scheme (#5889 )

2025-10-12 17:55:59 -07:00

test_session_control.py

Tiny move files to utils folder (#11166 )

2025-10-03 22:40:06 +08:00

test_skip_tokenizer_init.py

Support precomputed multimodal features for Qwen-VL and Gemma3 models. (#6136 )

2025-05-16 12:26:15 -07:00

test_srt_endpoint.py

[Feature] Add /tokenize and /detokenize OpenAI compatible endpoints (#9545 )

2025-10-08 12:38:48 +08:00

test_srt_engine_with_quant_args.py

Revert "fix some typos" (#6244 )

2025-05-12 12:53:26 -07:00

test_srt_engine.py

Tiny move files to utils folder (#11166 )

2025-10-03 22:40:06 +08:00

test_standalone_speculative_decoding.py

Move deep gemm related arguments to sglang.srt.environ (#11547 )

2025-10-14 00:34:35 +08:00

test_start_profile.py

Fix [test]: Env:SGLANG_TORCH_PROFILER_DIR for pytest. (#10780 )

2025-10-03 22:59:32 +08:00

test_swa_unittest.py

EAGLE cache fix for SWARadixCache (#11231 )

2025-10-07 18:21:37 +08:00

test_tokenizer_batch_encode.py

Tiny move files to utils folder (#11166 )

2025-10-03 22:40:06 +08:00

test_tokenizer_manager.py

Tiny move files to utils folder (#11166 )

2025-10-03 22:40:06 +08:00

test_torch_compile_moe.py

[UT for RL] Add UT to cover release/resume memory case for moe model (#8803 )

2025-09-09 19:25:12 -07:00

test_torch_compile.py

Improve profiler and integrate profiler in bench_one_batch_server (#6787 )

2025-05-31 15:53:55 -07:00

test_torch_flex_attention_backend.py

feat: Add FlexAttention Backend for Efficient Sparse Attention (#9947 )

2025-09-18 11:49:17 -07:00

test_torch_native_attention_backend.py

[CI] test chunked prefill more (#5798 )

2025-04-28 10:57:17 -07:00

test_torch_tp.py

[CI] test chunked prefill more (#5798 )

2025-04-28 10:57:17 -07:00

test_torchao.py

Replace time.time() to time.perf_counter() for benchmarking. (#6178 )

2025-05-11 14:32:49 -07:00

test_tracing.py

[Feature] Sglang Tracing: Fine-Grained Tracking for Request Latency - Part 1 (#9962 )

2025-09-15 02:08:02 +08:00

test_triton_attention_backend.py

[CI] test chunked prefill more (#5798 )

2025-04-28 10:57:17 -07:00

test_triton_attention_kernels.py

Deterministic Mode: Add 1-stage triton kernel for prefill (#11147 )

2025-10-20 01:47:36 +08:00

test_triton_attention_rocm_mla.py

Add retry for flaky tests in CI (#4755 )

2025-03-25 16:53:12 -07:00

test_triton_fused_moe.py

Refactor Triton-kernel MoE runner integration (#11795 )

2025-10-23 18:47:28 -07:00

test_triton_moe_channel_fp8_kernel.py

[6/N] MoE Refactor: Cleanup MoE-related configs (#8849 )

2025-08-14 21:14:53 -07:00

test_triton_moe_wna16.py

[6/N] MoE Refactor: Cleanup MoE-related configs (#8849 )

2025-08-14 21:14:53 -07:00

test_triton_sliding_window.py

Fix triton sliding window test case (#6981 )

2025-06-08 17:20:46 -07:00

test_two_batch_overlap.py

Move deep gemm related arguments to sglang.srt.environ (#11547 )

2025-10-14 00:34:35 +08:00

test_utils_update_weights.py

Fix CI (#9012 )

2025-08-09 13:33:42 -07:00

test_vertex_endpoint.py

Properly return error response in vertex_generate HTTP endpoint (#5956 )

2025-05-01 11:48:58 -07:00

test_vision_chunked_prefill.py

vlm: enforce pybase64 for image and str encode/decode (#10700 )

2025-10-21 19:05:32 +08:00

test_vision_openai_server_a.py

model: support deepseek-ocr (#11891 )

2025-10-24 03:15:17 +08:00

test_vision_openai_server_common.py

model: support deepseek-ocr (#11891 )

2025-10-24 03:15:17 +08:00

test_vllm_dependency.py

ci: refactor nightly test (#10495 )

2025-09-26 15:24:30 -07:00

test_vlm_accuracy.py

Move parsers under a single folder (#9912 )

2025-09-02 18:25:04 -07:00

test_vlm_input_format.py

Update transformers package version to 4.57.0 (#11222 )

2025-10-05 23:45:14 +00:00

test_wave_attention_backend.py

[AMD] Support Wave attention backend with AMD GPU optimizations (#8660 )

2025-08-12 13:49:11 -07:00

test_wave_attention_kernels.py

[AMD] Support Wave attention backend with AMD GPU optimizations (#8660 )

2025-08-12 13:49:11 -07:00

test_weight_version.py

feat: Add model version tracking with API endpoints and response metadata (#8795 )

2025-08-14 12:13:46 -07:00